High propagating velocity of spin waves and temperature dependent
damping in a CoFeB thin film
Haiming Yu,
a),b)
R. Huber,
a)
T. Schwarze, F. Brandl, T. Rapp, P. Berberich, G. Duerr,
and D. Grundler
c)
Lehrstuhl fu¨r Physik funktionaler Schichtsysteme, Technische Universita¨t Mu¨nchen, Physik Department,
James-Franck-Str. 1, D-85747 Garching b. Mu¨nchen, Germany
(Received 24 April 2012; accepted 9 June 2012; published online 29 June 2012)
Spin wave propagation in a magnetron-sputtered CoFeB thin film is investigated. We apply both
in-plane and out-of-plane magnetic fields. At room temperature, we find velocities of up to 25 and
3.5 km/s, respectively. These values are much larger compared to a thin permalloy film. Analyzing
the resonance linewidth, we obtain an intrinsic Gilbert damping parameter of about 0.007 at room
temperature. It increases to 0.023 at 5 K. CoFeB is a promising material for magnonic devices
supporting fast propagating spin waves.
V
C
2012 American Institute of Physics.
[http://dx.doi.org/10.1063/1.4731273]
The alloy CoFeB is used in magnetic tunnel junctions
(MTJs), which form the basis for magnetic random access
memory (MRAM), read heads in hard disk drives, as well as
spin-logic based devices.
1–3
At the same time, CoFeB films
have been relevant to explore photo-induced spin-wave
excitations in magnetic antidot lattices.
4
Ferromagnetic reso-
nance measurements evidenced a small linewidth for
CoFeB.
5–10
For magnonic applications,
11,12
electrically gen-
erated spin waves would be very interesting. Spin-transfer
torque (STT) which is found to occur in dc-current biased
CoFeB-based MTJs (Ref. 13) has been shown to exci te spin
dynamics.
14,15
It has been predicted that the excitation of
propagating spin waves by STT depends, however, crucially
on the out-of-plane component of the magnetization vector
M of the underlying film.
16–18
Experimental evidence has
been provided that over a critical angle a propagating spin
wave mode can be excited, and below this angle, both local-
ized and propagating spin waves are excited.
19
Madami et al.
recently observed a propagating spin wave induced by an
STT oscillator where they used a Ni
80
Fe
20
(permalloy) film
subject to an out-of-plane oriented magnetic field.
20
It is now
timely to explore spin-wave propagation velocities and
damping of CoFeB in out-of-plane magnetic fields and com-
pare them to permalloy which, so far, is mainly used for
magnonic nanod evices and magnonic crystals.
11,22,23
In this
paper, we present temperature-dependent measurements on
spin wave propagation and damping in CoFeB, applying,
both, in in-plane and out-of-plane magnetic fields. We
address the temperature range from 5 to 295 K. The intrinsic
damping parameter a
i
is found to be as small as 0.007 at
295 K. At the same time, the group velocity is measured to
be a factor of about 3 larger compared to permalloy provok-
ing a large decay for CoFeB. The damping parameter is
found to increase only below 200 K. We do not observe an
in-plane magnetic anisotropy making magnetron-sputtered
CoFeB an interesting material for magnonic devices with
electrical spin-wave injection.
Thin films were prepared on semi-insulating GaAs sub-
strates by magnetron sputtering in argon atmosphere using a
Co
20
Fe
60
B
20
target.
21
Here, we report data obtained on a
60 nm thick film forming a mesa with an area of 300 lm
120 lm as depicted in Fig. 1. A 5 nm thick layer of Al
2
O
3
was grown by atomic layer deposition onto the mesa to
ensure an electrical isolation for two shortened metallic co-
planar waveguides (CPWs) integrated by optical lithography.
These two CPWs were designed to act as a spin wave emitter
(CPW1) and detector (CPW2). The six conducting lines
were 2 lm wide. The edge-to-edge separation between signal
and ground lines was 1:6 lm. The distance s between the two
inner conductors was 12 lm. Reference permalloy (Py) thin
films were deposited using electron beam evaporation. The
thickness was 40 nm. To measure spin-wave propagation and
extract the Gilbert damping parameter a
i
, we connected a
vector network analyzer to the CPWs. We measured trans-
mission and reflection signals, respectively.
24
The Fourier
analysis of the microwave field generated around the inner
conductor of CPW1 provided us with an excitation spec-
trum
25
which contained two peaks I and II at wave vectors k
I
and k
II
, respectively. The wave vectors were perpendicular
to the CPW. The magnetic field H was applied, both, in the
film plane and in the out-of-plane direction.
FIG. 1. Microscopy image of a CoFeB mesa with two integrated coplanar
waveguides (CPWs). The center-to-center separation s of the two CPWs is
12 lm. Orientations of in-plane magnetic field H and wave vector k are
sketched.
a)
H. Yu and R. Huber contributed equally to this work.
b)
haiming.yu@ph.tum.de.
c)
grundler@ph.tum.de.
0003-6951/2012/100(26)/262412/4/$30.00
V
C
2012 American Institute of Physics100, 262412-1
APPLIED PHYSICS LETTERS 100, 262412 (2012)
Downloaded 29 Jun 2012 to 128.178.183.92. Redistribution subject to AIP license or copyright; see http://apl.aip.org/about/rights_and_permissions
Fig. 2(a) shows color-coded spectra obtained in trans-
mission configuration between emitter and detector CPWs
on CoFeB. We depict the imaginary part of the transmission
signal. H is collinear with the CPWs. We therefore excite
Damon-Eshbach-type (DE) modes. We observe two
branches which vary characteristically with H. Each of the
modes I and II generates a black-white-black oscillating con-
trast that indicates spin wave propagation. In Fig. 2(b),we
show a single spectrum extracted at 20 mT. Following Refs.
24 and 26, the frequency separation Df reflects a change in
the phase by 2p which is accumulated by spin waves propa-
gating between the inner conductors of CPW1 and CPW2.
The group velocity v
g
is calculated from Df according to
Ref. 24,
v
g
¼
@x
@k
2pDf
2p=s
¼ Df s: (1)
Using Eq. (1), we evaluate group velocities close to the rele-
vant wave vector k
I
. In Fig. 3(a), we summarize the veloc-
ities for DE modes in CoFeB as a function of H (filled
circles). At l
0
jHj¼6 mT, we find 25 km/s. The value
decreases with increasing absol ute value of l
0
jHj, reaching
15 km/s at 90 mT. Velocities measured on the permalloy ref-
erence film (open circles) are found to be a factor of 2.5 to 3
smaller compared to CoFeB. This is attributed mainly to the
different values of the saturation magnetization M
S
as will
be discussed later.
The excitation spectrum of the emitter CPW consists of
two peaks I and II reflecting the wave vectors k
I
¼ 0:6
10
4
rad=cm and k
II
¼ 2:4 10
4
rad=cm.
24
This gives rise
to the two distinct resonances I and II in the spectra (Fig. 2).
From the frequency separation (f
II
f
I
) between two peaks II
and I [c.f. Fig. 4(b)], a group velocity can be recalculated
according to v
g
¼ 2ðf
II
f
I
Þ=ðk
II
k
I
Þ in an independent
way. At 295 K and 20 mT, we obtain 18.6 km/s for CoFeB.
This value is slightly smaller compared to 22.3 km/s which
we evaluate from Eq. (1) in Fig. 3(b) at 20 mT and b ¼ 0
.
The discrepancy can be understood as follow. When using
peaks I and II for the evaluation, we average the group veloc-
ity over a broad wave vector range ðk
II
k
I
Þ. Since the group
velocity in general decreases with increasing k, the value
2ðf
II
f
I
Þ=ðk
II
k
I
Þ is expected to be smaller compared to
the slope v
g
¼ @x=@k ¼ 2p@f =@k used in Eq. (1) where we
evaluate v
g
close to k
I
only.
When we now fix l
0
jHj at 20 mT and rotate the field in
the plane of the ferromagnet, the velocity decreases with
increasing angle b as shown in Fig. 3(b). Here, b is the angle
between H and the CPW (Fig. 1). For CoFeB (filled circles),
the velocity drops from 22.3 to 9.0 km/s when b varies from
0to60
. The variation observed for CoFeB is more pro-
nounced compared to permalloy (open circles) where the ve-
locity varies from 8 to 4.5 km/s. We attribute the observed
angular dependence of the velocity to the change in angle
between the wave vector k and the magnetization vector M
which follows the rotation of H. We note that in separate res-
onance measurements (not shown) where we rotate the sam-
ple and keep the angle between k and H constantly at 90
,
we do not observe an in-plane magnetic anisotropy in, both,
the polycrystalline magnetron-sputtered CoFeB and evapo-
rated permalloy.
It is now interesting to study the propagation character-
istics with the field H applied in a direction perpendicular to
the plane. When l
0
jHj exceeds the shape anisotropy field of
CoFeB of about 1.7 T, we obtain transmission spectra such
as the one shown in Fig. 4(a). The oscillating signal marked
by Df now reflects propagating magnetostatic forward-
FIG. 2. (a) Gray-scale plot of spectra measured on CoFeB at room tempera-
ture in transmission configuration between two neighboring coplanar wave-
guides. The field is collinear with the CPWs. The dashed line indicates
20 mT where we extract the spectrum shown in (b). The spectrum in (b) is
the imaginary part of the transmission signal. The frequency separation
denoted by Df is used to calculate the group velocity. The emitter CPW
excites two pronounced modes I and II.
FIG. 3. (a) Group velocities of DE modes at room temperature as a function
of applied field. Filled (open) circles denote CoFeB (permalloy). (b) Spin
wave group velocity as a function of angle b when the in-plane field
amounts to l
0
H ¼ 20 mT.
FIG. 4. (a) Imaginary part of the transmission signal obtained on CoFeB
when the applied field of 1.9 T is perpendicular to the film plane. Df indi-
cates the phase shift of 2p acquired by the MSFVWs. (b) Imaginary part of
the reflection signal with an applied field of 1.9 T being perpendicular to the
film plane. Both spectra were measured at 5 K.
262412-2 Yu et al. Appl. Phys. Lett. 100, 262412 (2012)
Downloaded 29 Jun 2012 to 128.178.183.92. Redistribution subject to AIP license or copyright; see http://apl.aip.org/about/rights_and_permissions
volume waves (MSFVWs). The group velocities evaluated
between 5 and 295 K are summarized in Fig. 5(a). At 295 K
the MSFVW exhibits a velocity of 3.5 km/s. This value is a
factor of seven smaller compared to the DE mode. The ve-
locity increases by 25% when cooling down to 5 K, reaching
4.3 km/s. The corresponding values of the permalloy thin
film are 1.2 (295 K) and 1.5 km/s (5 K). Again, the velocities
measured on permalloy are smaller by about a factor of
three.
We measured spectra also in reflection configuration
[Fig. 4(b)] to extract information on the linewidth df and
damping in the perpendicular-to-plane field configuration.
For this, we fitted a Lorentz curve to mode I and extracted
the full width at half maximum as df . We find df to increase
almost linearly with f at a given temperature [inset of
Fig. 5(b)]. The eigenfrequency f is varied by changing the
magnetic field H. We analyzed the slope df vs f using,
27
df ¼
jcj
2p
l
0
DH þ 2a
i
f ; (2)
where c is the gyromagnetic splitting factor, and DH reflects
a line broadening caused by extrinsic mechanisms and film
inhomogeneities. Following Eq. (2), the slope in the inset of
Fig. 5(b) provides us with the intrinsic damping parameter
a
i
. We have extracted the damping of the CoFeB thin film in
perpendicular magnetic fields at different temperatures T and
summarize the values in Fig. 5(b). For T 200 K, we get
0.007 consistent with the earlier reports on CoFeB alloys.
Down to 5 K the damping increases by a factor of three to a
value of 0.023. We have used the ratio between the signal in-
tensity at the detector and the reflected signal intensity at the
emitter I
21
=I
22
to estimate the decay lengths.
24,31
At room
temperature, we obtain a value of 23.9 lm for the DE mode
and 4.6 lm for the MSFVW mode in CoFeB.
We now discuss our results. The propagating velocities
measured on CoFeB and permalloy are found to differ by
about a factor of three. This factor was observed for, both,
DE modes and MSFVWs. In the append ix of Ref. 28, group
velocities are evaluated for different orientations of M and k
as a function of film thickness d and saturation magnetization
M
S
. For DE modes, the group velocity reads
1
v
g
¼
4
x
M
d
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
x
0
ðx
0
þ x
M
Þ
p
x
M
: (3)
For MSFVW one gets
1
v
g
¼
4
x
M
d
: (4)
In Eqs. (3) and (4), x
0
¼cl
0
H, and x
M
¼cl
0
M
S
. For
CoFeB and permalloy, we assume values of l
0
M
S
¼ 1.8(2) T
(Ref. 4) and 1.01 T,
29
respectively. Considering DE modes at
an applied field of a few 10 mT, the external field is much
smaller than the saturation magnetization. In this limit, we
obtain v
g
¼
cl
0
d
4
H
1
2
M
3
2
S
from Eq. (3). Starting from a small
field H, the group velocity is expected to decrease with
increasing H according to H
1
2
. This explains why we detect
the largest propagation velocities near remanence in Fig.
3(a). Considering the different saturation magnetizations M
S
and thicknesses d, Eq. (3) predicts group velocities which
differ by a factor of
d
CoFeB
M
3=2
S;CoFeB
d
Py
M
3=2
S;Py
¼ 3:4 between CoFeB and
permalloy.
30
This is close to the factor of three observed in
Fig. 3(a). For MSFVWs, Eq. (4) predicts v
g
to be 4.8 km/s
for CoFeB. We find a value close to this at low temperature
in Fig. 5(a).
In general, microscopic origins of spin wav e damping
include direct and indirect contributions for angular momen-
tum transfer to the lattice.
32
For MSFVW, indirect damping
such as two-magnon scattering is suppressed due to the out-
of-plane orientation of the magnetization.
28,33
We thus at-
tribute the damping parameter in Fig. 5(b) to direct damping
where the spin angular momentum is transferred to non-
magnetic degrees of freedom. Processes involv ed are intra-
and inter-band electronic transitions provoked by spin
precession in the presence of spin-orbit interaction. Interest-
ingly, we find in Fig. 5(b) that the damping parameter
increases below a temperature of about 200 K. Such an
increase of a
i
with decreasing temperature has already been
observed for further metallic ferromagnets.
34
It is attributed
to the variation of the electron scattering time which
increases with decreasing T when electron scattering due to
phonons diminishes. A longer electron scattering time favors
angular momentum transfer due to direct damping via intra-
band transitions.
32
The observed decrease of the damping
constant with increasing temperature is consistent with the
theoretical work of Gilmore et al.,
35
who have used first-
principles calculations and discussed damping for ferromag-
netic elements such as Fe, Co, and Ni. The observed low
damping might be understood by the suppression of spin-flip
FIG. 5. (a) Group velocity measured on CoFeB in a perpendicular field of
1.9 T at different temperatures T. The eigenfrequency varies from 5.6 GHz
to 7.2 GHz when T increases from 5 to 295 K. (b) Temperature dependence
of intrinsic damping parameter a
i
. The inset shows df vs f(H) at 295 K. The
line reflects Eq. (2) leading to a
i
¼ 0.007.
262412-3 Yu et al. Appl. Phys. Lett. 100, 262412 (2012)
Downloaded 29 Jun 2012 to 128.178.183.92. Redistribution subject to AIP license or copyright; see http://apl.aip.org/about/rights_and_permissions
transitions which has been reported to occur for half metals
and is attributed also to pseudo-gap materials with high spin
polarization to which CoFeB alloys belong.
36
A detailed anal-
ysis of the resonance peak I in Fig. 4(b) provides that the peak
contains a further resonance on the low frequency side. With
increasing temperature this additional resonance disappears,
providing a Lorentzian-shaped peak at room temperature
when the damping is found to be smallest. So far the origin of
the additional feature at low T is not clear.
In conclusion, we have investigated group velocities of
spin waves in CoFeB. We obtain values significantly in
excess of permalloy which we attribute to the relatively large
saturation magnetization of CoFeB. Large decay lengths are
to be expected at room temperature as the damping is found
to be small and only 0.007.
Financial support by the German Excellence Cluster
Nanosystems Initiative Munich (NIM) and the European
Community’s Seventh Framework Programme (FP7/2007-
2013) under Grant Agreement No. 228673 MAGNONICS
and the DFG via project GR1640/5-1 in the priority pro-
gramme SPP1538 is gratefully acknowledged.
1
J. S. Moodera, L. R. Kinder, T. M. Wong, and R. Meservey, Phys. Rev.
Lett. 74, 3273 (1995).
2
S. S. P. Parkin, C. Kaiser, A. Penchula, P. M. Rice, B. Hughes, M. Samant,
and S. H. Yang, Nature Mater. 3, 862 (2004).
3
S. Ikeda, K. Miura, H. Yamamoto, K. Mizunuma, H. D. Gan, M. Endo, S.
Kanai, J. Hayakawa, F. Matsukura, and H. Ohno, Nature Mater. 9, 721
(2010).
4
H. Ulrichs, B. Lenk, and M. Mu¨nzenberg, Appl. Phys. Lett. 97, 092506
(2010).
5
C. Bilzer, T. Devolder, J. V. Kim, G. Counil, C. Chappert, S. Cardoso, and
P. P. Freitas, J. Appl. Phys. 100, 053903 (2006).
6
G. Malinowski, K. C. Kuiper, R. Lavrijsen, H. J. M. Swagten, and B.
Koopmans, Appl. Phys. Lett. 98, 153107 (2011).
7
H. Lee, L. Wen, M. Pathak, P. Janssen, P. Leclair, C. Alexander, C. K. A.
Mews, and T. Mewes, J. Phys. D: Appl. Phys. 41, 215001 (2008).
8
M. Oogane, T. Wakitani, S. Yakata, R. Yilgin, Y. Ando, A. Sakuma, and
T. Miyazaki, Jpn. J. Appl. Phys. 45, 3889 (2005).
9
X. Liu, W. Z. Matthew, J. Carter, and G. Xiao, J. Appl. Phys. 110, 033910
(2011).
10
F. Xu, Q. Huang, Z. Liao, S. Li, Y. Ando, C. K. Ong, and T. Miyazaki,
J. Appl. Phys. 111, 07A304 (2012).
11
S. Neusser and D. Grundler, Adv. Mater. 21, 2927 (2932).
12
V. V. Kruglyak, S. O. Demokritov, and D. Grundler, J. Phys. D: Appl.
Phys. 43, 264001 (2010).
13
D. C. Worledge, G. Hu, D. W. Abraham, J. Z. Sun, P. L. Trouilloud, J.
Nowak, S. Brown, M. C. Gaidis, E. J. OSullivan, and R. P. Robertazzi,
Appl. Phys. Lett. 98, 022501 (2011).
14
S. I. Kiselev, J. C. Sankey, I. N. Krivorotov, N. C. Emley, R. J. Schoel-
kopf, R. A. Buhrman, and D. C. Ralph, Nature 425, 380 (2003).
15
W. H. Rippard, M. R. Pufall, S. Kaka, S. E. Russek, and T. J. Silva, Phys.
Rev. Lett. 92, 027201 (2004).
16
G. Gerhart, E. Bankowski, G. A. Melkov, V. S. Tiberkevich, and A. N.
Slavin, Phys. Rev. B 76, 024437 (2007).
17
G. Consolo, B. Azzerboni, L. Lopez-Diaz, G. Gerhart, E. Bankowski, V.
Tiberkevich, and A. Slavin, Phys. Rev. B 78, 014420 (2008).
18
M. A. Silva, T. J. Silva, and M. D. Stiles, Phys. Rev. B 77, 144401 (2008).
19
S. Bonetti, V. Tiberkevich, G. Consolo, G. Finocchio, P. Muduli, F. Manc-
off, A. Slavin, and J. Akerman, Phys. Rev. Lett. 105, 217204 (2010).
20
M. Madami, S. Bonetti, G. Consolo, S. Tacchi, G. Carlotti, G. Gubbiotti,
F. B. Mancoff, and J. A
˚
kerman, Nature Nanotechnol. 10, 1038 (2011).
21
Using energy dispersive x-ray analysis of a nominally identical thin film,
we find that the composition of the thin film differs from the target compo-
sition. For the atomic concentration ratio between Fe and Co, we find 3.6
instead of 3. The exact relative amount of B is not known at this point.
22
V. E. Demidov, S. Urazhdin, E. R. J. Edwards, M. D. Stiles, R. D. McMi-
chael, and S. O. Demokritov, Phys. Rev. Lett. 107, 107204 (2011).
23
J. Topp, D. Heitmann, and D. Grundler, Phys. Rev. B 80, 174421 (2009).
24
S. Neusser, G. Duerr, H. G. Bauer, S. Tacchi, M. Madami, G. Woltersdorf,
G. Gubbiotti, C. H. Back, and D. Grundler, Phys. Rev. Lett. 105, 067208
(2010).
25
G. E. Ponchak, E. M. Tentzeris, and L. P. B. Katehi, Int. J. Microcircuits
Electron. Packag. 20, 167 (1997).
26
V. Vlaminck and M. Bailleul, Science 322, 410 (2008).
27
B. Kuanr, R. E. Camley, and Z. Celinski, Appl. Phys. Lett. 87(1), 012502
(2005).
28
D. D. Stancil and A. Prabhakar, Spin Waves: Theory and Applications,
Appendix C (Springer, 2009).
29
G. Duerr, M. Madami, S. Neusser, S. Tacchi, G. Gubbiotti, G. Carlotti,
and D. Grundler, Appl. Phys. Lett. 99, 202502 (2011).
30
We consider only the relative factor. Absolute values given by Eq. (3) dif-
fer by about a factor of 2 from the measured ones.
31
M. Bailleul, D. Olligs, and C. Fermon, Appl. Phys. Lett. 83, 972 (2003).
32
M. Fa¨hnle and C. Illg, J. Phys.: Condens. Matter 23, 493201 (2011).
33
R. Arias and D. L. Mills, Phys. Rev. B 60, 7395 (1999).
34
G. Counil, T. Devolder, J. V. Kim, P. Crozat, C. Chappert, S. Zoll, and R.
Fournel, IEEE Trans. Magn. 42, 3323 (2006).
35
K. Gilmore, Y. U. Idzerda, and M. D. Stiles, Phys. Rev. Lett. 99, 027204
(2007).
36
A. Mann, J. Walowski, M. Mu¨nzenberg, S. Maat, M. J. Carey, J. R. Childr-
ess, C. Mewes, D. Ebke, V. Drewello, G. Reiss, and A. Thomas,
arXiv:1202.3874v1 [cond-mat.mtrl-sci].
262412-4 Yu et al. Appl. Phys. Lett. 100, 262412 (2012)
Downloaded 29 Jun 2012 to 128.178.183.92. Redistribution subject to AIP license or copyright; see http://apl.aip.org/about/rights_and_permissions
